I use an Intel Xeon E5 1620 v2 CPU and a Nvidia Tesla GPU as devices.
I use the “intel sdk for opencl 2016 ubuntu 6.0.0.1049” to use the CPU as OpenCL device and “cuda_7.5.18_linux” to use the GPU as OpenCL device.
The host program runs on the CPU.
I created an event and use it to check if the enqueueReadBuffer has finished.
cl::Event readResult;
_command_queue.enqueueReadBuffer( _DevBufOut, CL_FALSE, 0, CHUNK_SIZE_1 * sizeof( float ), _cDataOut, NULL, &readResult);
readResult.wait();
I would like to know if the function cl::Event.wait() uses busy waiting.